<MasterR3C0RD>
apritzel: No expectations here of working on older kernels for many reasons, most important being that RTC is critical for bringing up the xradio WiFi module as it's clocked by fanout
<MasterR3C0RD>
Secondly because Ethernet isn't functional (I've been busy and haven't had the time to clean up my patches to submit), so zero networking and zero display means useless without the upcoming patches
<apritzel>
I see, thanks. In general we would need more board DT files in the kernel repo, so far there is only this AW internal devboard, which nobody has ever seen or tested
<apritzel>
I have a halfway working .dts for the Liontron board, though am missing some regulators still
<apritzel>
it's a bit shocking that there is basically zero documentation or not even an update image for this board (given it's a proper development board)
<apritzel>
mach
<MasterR3C0RD>
I'm not confident at all about the regulator setup for my Sonic Pad; even though I have the actual DTS, half of the regulators (including ones that are obviously critical) are set to fixed voltages, others are assigned to the wrong devices... it's an absolute mess
<apritzel>
MasterR3C0RD: if you could prepare a .dts for your board, that would be nice, we need some existing devices in the kernel repo
<apritzel>
yeah, the vendor DTS is basically useless when it comes to the regulator setup or assignment
<apritzel>
I typically try to dump the regulator setup from a running BSP system, that's the most reliable info in my experience
<apritzel>
and then double checking that by sabotaging rails and turning them off at runtime, and seeing what breaks ;-)
<MasterR3C0RD>
Dumping the regulators didn't really help much for me; I did a lot of "turn off regulator, turn on regulator" in U-Boot to figure out what powered what, but there's still some critical rails I don't quite know what they're powering
<apritzel>
well, if you can show that they are critical, that's good enough to mark them as always-on, I'd say
<apritzel>
sometimes we can tell by the voltage and other typical assignments, on similar boards or on boards using the same PMIC
<MasterR3C0RD>
I'll post my current DTS somewhere if you want to take a look and see if anything pops out; I did leave some comments in there regarding some of the lines
<MasterR3C0RD>
It's a bit of a mess of random settings in some spots, and technically still need to model that vcc_5v actually comes from a 12V DC power plug and maybe goes through some other regulators to create the 5V used by basically everything except the display I think, but really just needed something that works
<apritzel>
yeah, fair enough, we can iron out those problems on the list. Just wanted to point of that we only have sun50i-a100-allwinner-perf1.dts in mainline atm
<MasterR3C0RD>
At the very least I know that parthiban's DE series (whenever he's able to properly submit it) will have a DTS for the Helperboard
<loki666>
oh my god... what am going to do with my A133P ...
<apritzel>
loki666: yours is/will be one of those portable retro gaming devices?
<loki666>
yup
<apritzel>
probably one of the more popular and available A133 devices then, counting on you for a .dts file ;-)
<MasterR3C0RD>
loki666: You'll be at the forefront of making the A100/A133/A133P/R818/Creality T800 work on mainline :)
<MasterR3C0RD>
So many bins/rebadges...
<loki666>
yeah... I'll need help, and I really laks the tools to understand WTF is going on
<loki666>
even my multimeter has too big probes
<loki666>
apritzel: the usb oth patch has a big issue, as it calls phy_power_on/power_off from the detection job, we are getting some weird race conditions issues
<loki666>
something else is calling these ops while the job is running them
<loki666>
for the A133P i have acces to a tina bsp (haven't dl yet)
<loki666>
the thing is 10GB!
<MasterR3C0RD>
loki666: Send that my way if you can
<MasterR3C0RD>
loki666: For reference, that's the first BSP I've seen for an A133P device, which is quite special
<MasterR3C0RD>
Though I find amusing that in the GitHub link they stated "we have to see if we can release the sources first". Hopefully someone will let them know about the GPL
<loki666>
lol
<MasterR3C0RD>
Oh my god... it's a zip of tar files
<apritzel>
yeah, those BSPs are total carnage, typically: they might include a binary toolchain, a lot of build scripts, unrelated userland stuff, source files marked as executable, with DOS line endings, etc.
<MasterR3C0RD>
However, interesting find from `sys_config.fex`: `dram_type = 4` under [dram_para], matched with `select_mode = 0` under [dram_select_para]
<MasterR3C0RD>
Which means this might be another DDR4 device
<MasterR3C0RD>
apritzel: Yeah, it's a standard lichee toolchain with hundreds of packages in `dl/`, seems to have precompiled boot0 though unlike the nice libdram.o-sort of thing I found for the Sonic Pad
<MasterR3C0RD>
AXP2202 regulator, DSI display (which is tested and working with parthiban's series to enable the DE)
<MasterR3C0RD>
But looks like it also has the same TCS4838 regulator thing as my A133P
apritzel has quit [Ping timeout: 480 seconds]
Daanct12 has joined #linux-sunxi
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
hexdump01 has joined #linux-sunxi
hexdump0815 has quit [Ping timeout: 480 seconds]
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
Daanct12 has quit [Ping timeout: 480 seconds]
hexdump01 has quit [Read error: Connection reset by peer]
Schimsalabim has quit [Read error: Connection reset by peer]
hexdump01 has joined #linux-sunxi
Schimsalabim has joined #linux-sunxi
Daanct12 has joined #linux-sunxi
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
nickaa has joined #linux-sunxi
nickaa has quit []
apritzel has joined #linux-sunxi
sh1 has quit [Ping timeout: 480 seconds]
sh1 has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
sh1 has quit [Read error: Connection reset by peer]
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
gsz has joined #linux-sunxi
dsimic is now known as Guest5878
dsimic has joined #linux-sunxi
Guest5878 has quit [Ping timeout: 480 seconds]
<montjoie>
a bit unrelated but I seek people with 4 USB serial to confirm the weirdest linux-next fail I found
<montjoie>
(related to sunxi, since it fail my sunxi CI)
asriel has quit [Quit: Don't drink the water. They put something in it to make you forget.]
Danct12 has joined #linux-sunxi
<hlauer>
I saw an USB stuck on a banana M1+ last week which was only resolvable by a reboot - but you mean something else?
asriel has joined #linux-sunxi
<montjoie>
yes and finally 4 serial is not enough to trigger the issue :(
<apritzel>
montjoie: what do you mean exactly? A host (whatever device, x86 or arm) running linux-next, and connecting four USB-to-serial adapters? Or one USB-to-quad-serial adapter, like this FTDI4232?
<wigyori>
cut-down version of v831 or something else?
JohnDoe_71Rus has joined #linux-sunxi
digetx is now known as Guest5903
digetx has joined #linux-sunxi
Guest5903 has quit [Ping timeout: 480 seconds]
<loki666>
MasterR3C0RD any DTS?
junari has joined #linux-sunxi
Daanct12 has quit [Quit: WeeChat 4.4.4]
junari has quit [Quit: Leaving]
junari has joined #linux-sunxi
JohnDoe2 has joined #linux-sunxi
JohnDoe_71Rus has quit [Ping timeout: 480 seconds]
JohnDoe2 has quit [Ping timeout: 480 seconds]
JohnDoe2 has joined #linux-sunxi
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
<MasterR3C0RD>
loki666: For your device? It's in there, but it'd only work with that specific vendor kernel
<loki666>
Ok my unit should be there by tomorrow... I'll start with a buildroot environment... Hopefully I won't have to solder for the UART
<apritzel>
loki666: I am afraid you have to start from scratch with your DT, the vendor DTs are completely unusable for mainline
<loki666>
I guess I should first try to extract as much as possible info from the BSP... Hacking an init script to setup a ssh demon will be my first step
<loki666>
What should I be looking for ?
<apritzel>
the most interesting is the regulator assignment, so if you are root, then dump /sys/kernel/debug/regulator/regulator_summary
<apritzel>
also /sys/kernel/debug/gpio, to learn about enable or IRQ GPIOs
<apritzel>
there *is* some information in the DT, but you have to extract it from there, and put it in your DT, in a proper way
<apritzel>
since there is no good A100/A133 DT in the tree, I'd suggest you look at the H616 ones instead, for instance for the Anbernic devices
<loki666>
ok thanks
<MasterR3C0RD>
loki666: Note that you should not try to include `opp` on your device, as the DC-DC regulators used for the CPU on your board don't have a driver in mainline. Thus you'll be stuck with lower clocks for the time being
<loki666>
Ok good to know... How is u-boot?
<MasterR3C0RD>
You'll need to do some manual work, but considering your device appears on first glance to have DDR4, you might have an easier time. You'll need UART to get proper parameters though
<loki666>
Yeah that was clear
<apritzel>
well, you can maybe more reliably extract the DRAM parameters from some (update) image file, using sunxi-fw
<MasterR3C0RD>
Hmm, figured out what company made the TCS4838 at least (Torch-Chip Semiconductor), but no luck finding any sort of datasheet
gnarface has quit [Ping timeout: 480 seconds]
gnarface has joined #linux-sunxi
bauen1 has quit [Ping timeout: 480 seconds]
<apritzel>
MasterR3C0RD: yeah, same here: but there is code on github, in AW BSP U-Boot trees, which I think reveal enough info to write a driver
<apritzel>
since it's a rather simple single channel DC/DC converter, that can't be too hard
<apritzel>
I guess drivers/regulator/sy8106a-regulator.c could be used as a template, for a single channel I2C controlled regulator
apritzel has quit [Ping timeout: 480 seconds]
BroderTuck has joined #linux-sunxi
<loki666>
DC/DC reg is the AXP717 equivalent ?
Schimsalabim has quit [Read error: Connection reset by peer]
<BroderTuck>
Interesting. The Teclast_P85T page says that the first part of the SID for A523 is 02c05200 whereas on my proplus box (H728) it's 0300ff00- Very different.
<BroderTuck>
Whats the value on the (T527) Avaota board?
linkmauve has left #linux-sunxi [Error from remote client]
linkmauve has joined #linux-sunxi
vagrantc has joined #linux-sunxi
<MasterR3C0RD>
loki666: No; it's separate from the AXP2202 (AXP717) and acts to provide power solely to the ARM cores
<MasterR3C0RD>
The rest of the power rails to the A133P are still handled by an AXP PMIC, but my belief is that with the much higher clocks possible with the 2.0GHz clocks, more current is needed; thus why it needs a separate regulator
<MasterR3C0RD>
(more current is needed than can be provided by polyphased rails on the AXP chips)